INDEX

Explanations

words relating to actors or groups of people in a specific context

oai_token-act-pair · gpt-3.5-turbo

terms related to ensemble or supporting casts in various contexts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Fountain

-0.61

 saline

-0.60

éĽ

-0.59

ishable

-0.58

paren

-0.57

 Parenthood

-0.56

trade

-0.56

 Caption

-0.56

 accur

-0.55

verbs

-0.54

POSITIVE LOGITS

members

1.34

 members

1.14

 consisting

1.13

 comprising

1.12

 comprised

1.12

naire

1.11

 mates

1.09

 assembled

1.09

member

0.99

 comprises

0.98

Activations Density 0.300%

words relating to actors or groups of people in a specific context

terms related to ensemble or supporting casts in various contexts

No Comments

No Known Activations